The Agent Skills Directory

[SAFE]: No malicious behaviors or security risks were identified. The skill performs standard local file system operations including reading project metadata (e.g., package.json, go.mod) and creating a dedicated configuration directory (openspec/).- [PROMPT_INJECTION]: The skill reads external project files to detect the tech stack and architecture. While these files are untrusted inputs that are interpolated into a configuration file, the skill does not execute the ingested content, limiting the risk to metadata poisoning within the local config file.- [COMMAND_EXECUTION]: The skill performs directory creation and file writing. These operations are scoped to the project's root and are necessary for the skill's documented functionality.

sdd-init